1PING(8) System Manager's Manual: iputils PING(8)
2
6 ping, ping6 - send ICMP ECHO_REQUEST to network hosts
7
9 ping [ -LRUbdfnqrvVaAB] [ -c count] [ -m mark] [ -i interval] [ -l
10 preload] [ -p pattern] [ -s packetsize] [ -t ttl] [ -w deadline] [
11 -F flowlabel] [ -I interface] [ -M hint] [ -N nioption] [ -Q tos]
12 [ -S sndbuf] [ -T timestamp option] [ -W timeout] [ hop ...] desti‐
13 nation
14
17 ping uses the ICMP protocol's mandatory ECHO_REQUEST datagram to elicit
18 an ICMP ECHO_RESPONSE from a host or gateway. ECHO_REQUEST datagrams
19 (``pings'') have an IP and ICMP header, followed by a struct timeval
20 and then an arbitrary number of ``pad'' bytes used to fill out the
21 packet.
22 ping6 can also send Node Information Queries (RFC4620).
24
26 -a Audible ping.
27 -A Adaptive ping. Interpacket interval adapts to round-trip time,
29 so that effectively not more than one (or more, if preload is
30 set) unanswered probes present in the network. Minimal interval
31 is 200msec for not super-user. On networks with low rtt this
32 mode is essentially equivalent to flood mode.
33 -b Allow pinging a broadcast address.
35 -B Do not allow ping to change source address of probes. The
37 address is bound to one selected when ping starts.
38 -m mark
40 use mark to tag the packets going out. This is useful for vari‐
41 ety of reasons within the kernel such as using policy routing to
42 select specific outbound processing.
43 -c count
45 Stop after sending count ECHO_REQUEST packets. With deadline
46 option, ping waits for count ECHO_REPLY packets, until the time‐
47 out expires.
48 -d Set the SO_DEBUG option on the socket being used. Essentially,
50 this socket option is not used by Linux kernel.
51 -F flow label
53 Allocate and set 20 bit flow label on echo request packets.
54 (Only ping6). If value is zero, kernel allocates random flow
55 label.
56 -f Flood ping. For every ECHO_REQUEST sent a period ``.'' is
58 printed, while for ever ECHO_REPLY received a backspace is
59 printed. This provides a rapid display of how many packets are
60 being dropped. If interval is not given, it sets interval to
61 zero and outputs packets as fast as they come back or one hun‐
62 dred times per second, whichever is more. Only the super-user
63 may use this option with zero interval.
64 -i interval
66 Wait interval seconds between sending each packet. The default
67 is to wait for one second between each packet normally, or not
68 to wait in flood mode. Only super-user may set interval to val‐
69 ues less 0.2 seconds.
70 -I interface address
72 Set source address to specified interface address. Argument may
73 be numeric IP address or name of device. When pinging IPv6 link-
74 local address this option is required.
75 -l preload
77 If preload is specified, ping sends that many packets not wait‐
78 ing for reply. Only the super-user may select preload more than
79 3.
80 -L Suppress loopback of multicast packets. This flag only applies
82 if the ping destination is a multicast address.
83 -N nioption
85 Send ICMPv6 Node Information Queries (RFC4620), instead of Echo
86 Request.
87 name Queries for Node Names.
89 ipv6 Queries for IPv6 Addresses. There are several IPv6 spe‐
91 cific flags.
92 ipv6-global
94 Request IPv6 global-scope addresses.
95 ipv6-sitelocal
97 Request IPv6 site-local addresses.
98 ipv6-linklocal
100 Request IPv6 link-local addresses.
101 ipv6-all
103 Request IPv6 addresses on other interfaces.
104 ipv4 Queries for IPv4 Addresses. There is one IPv4 specific
106 flag.
107 ipv4-all
109 Request IPv4 addresses on other interfaces.
110 subject-ipv6=ipv6addr
112 IPv6 subject address.
113 subject-ipv4=ipv4addr
115 IPv4 subject address.
116 subject-name=nodename
118 Subject name. If it contains more than one dot, fully-
119 qualified domain name is assumed.
120 subject-fqdn=nodename
122 Subject name. Fully-qualified domain name is always
123 assumed.
124 -n Numeric output only. No attempt will be made to lookup symbolic
126 names for host addresses.
127 -p pattern
129 You may specify up to 16 ``pad'' bytes to fill out the packet
130 you send. This is useful for diagnosing data-dependent problems
131 in a network. For example, -p ff will cause the sent packet to
132 be filled with all ones.
133 -D Print timestamp (unix time + microseconds as in gettimeofday)
135 before each line.
136 -Q tos Set Quality of Service -related bits in ICMP datagrams. tos can
138 be either decimal or hex number. Traditionally (RFC1349), these
139 have been interpreted as: 0 for reserved (currently being rede‐
140 fined as congestion control), 1-4 for Type of Service and 5-7
141 for Precedence. Possible settings for Type of Service are: min‐
142 imal cost: 0x02, reliability: 0x04, throughput: 0x08, low delay:
143 0x10. Multiple TOS bits should not be set simultaneously. Pos‐
144 sible settings for special Precedence range from priority (0x20)
145 to net control (0xe0). You must be root (CAP_NET_ADMIN capabil‐
146 ity) to use Critical or higher precedence value. You cannot set
147 bit 0x01 (reserved) unless ECN has been enabled in the kernel.
148 In RFC2474, these fields has been redefined as 8-bit Differenti‐
149 ated Services (DS), consisting of: bits 0-1 of separate data
150 (ECN will be used, here), and bits 2-7 of Differentiated Ser‐
151 vices Codepoint (DSCP).
152 -q Quiet output. Nothing is displayed except the summary lines at
154 startup time and when finished.
155 -R Record route. Includes the RECORD_ROUTE option in the
157 ECHO_REQUEST packet and displays the route buffer on returned
158 packets. Note that the IP header is only large enough for nine
159 such routes. Many hosts ignore or discard this option.
160 -r Bypass the normal routing tables and send directly to a host on
162 an attached interface. If the host is not on a directly-
163 attached network, an error is returned. This option can be used
164 to ping a local host through an interface that has no route
165 through it provided the option -I is also used.
166 -s packetsize
168 Specifies the number of data bytes to be sent. The default is
169 56, which translates into 64 ICMP data bytes when combined with
170 the 8 bytes of ICMP header data.
171 -S sndbuf
173 Set socket sndbuf. If not specified, it is selected to buffer
174 not more than one packet.
175 -t ttl Set the IP Time to Live.
177 -T timestamp option
179 Set special IP timestamp options. timestamp option may be
180 either tsonly (only timestamps), tsandaddr (timestamps and
181 addresses) or tsprespec host1 [host2 [host3 [host4]]] (timestamp
182 prespecified hops).
183 -M hint
185 Select Path MTU Discovery strategy. hint may be either do (pro‐
186 hibit fragmentation, even local one), want (do PMTU discovery,
187 fragment locally when packet size is large), or dont (do not set
188 DF flag).
189 -U Print full user-to-user latency (the old behaviour). Normally
191 ping prints network round trip time, which can be different f.e.
192 due to DNS failures.
193 -v Verbose output.
195 -V Show version and exit.
197 -w deadline
199 Specify a timeout, in seconds, before ping exits regardless of
200 how many packets have been sent or received. In this case ping
201 does not stop after count packet are sent, it waits either for
202 deadline expire or until count probes are answered or for some
203 error notification from network.
204 -W timeout
206 Time to wait for a response, in seconds. The option affects only
207 timeout in absense of any responses, otherwise ping waits for
208 two RTTs.
209 When using ping for fault isolation, it should first be run on the
211 local host, to verify that the local network interface is up and run‐
212 ning. Then, hosts and gateways further and further away should be
213 ``pinged''. Round-trip times and packet loss statistics are computed.
214 If duplicate packets are received, they are not included in the packet
215 loss calculation, although the round trip time of these packets is used
216 in calculating the minimum/average/maximum round-trip time numbers.
217 When the specified number of packets have been sent (and received) or
218 if the program is terminated with a SIGINT, a brief summary is dis‐
219 played. Shorter current statistics can be obtained without termination
220 of process with signal SIGQUIT.
221 If ping does not receive any reply packets at all it will exit with
223 code 1. If a packet count and deadline are both specified, and fewer
224 than count packets are received by the time the deadline has arrived,
225 it will also exit with code 1. On other error it exits with code 2.
226 Otherwise it exits with code 0. This makes it possible to use the exit
227 code to see if a host is alive or not.
228 This program is intended for use in network testing, measurement and
230 management. Because of the load it can impose on the network, it is
231 unwise to use ping during normal operations or from automated scripts.
232
234 An IP header without options is 20 bytes. An ICMP ECHO_REQUEST packet
235 contains an additional 8 bytes worth of ICMP header followed by an
236 arbitrary amount of data. When a packetsize is given, this indicated
237 the size of this extra piece of data (the default is 56). Thus the
238 amount of data received inside of an IP packet of type ICMP ECHO_REPLY
239 will always be 8 bytes more than the requested data space (the ICMP
240 header).
241 If the data space is at least of size of struct timeval ping uses the
243 beginning bytes of this space to include a timestamp which it uses in
244 the computation of round trip times. If the data space is shorter, no
245 round trip times are given.
246
248 ping will report duplicate and damaged packets. Duplicate packets
249 should never occur, and seem to be caused by inappropriate link-level
250 retransmissions. Duplicates may occur in many situations and are
251 rarely (if ever) a good sign, although the presence of low levels of
252 duplicates may not always be cause for alarm.
253 Damaged packets are obviously serious cause for alarm and often indi‐
255 cate broken hardware somewhere in the ping packet's path (in the net‐
256 work or in the hosts).
257
259 The (inter)network layer should never treat packets differently depend‐
260 ing on the data contained in the data portion. Unfortunately, data-
261 dependent problems have been known to sneak into networks and remain
262 undetected for long periods of time. In many cases the particular pat‐
263 tern that will have problems is something that doesn't have sufficient
264 ``transitions'', such as all ones or all zeros, or a pattern right at
265 the edge, such as almost all zeros. It isn't necessarily enough to
266 specify a data pattern of all zeros (for example) on the command line
267 because the pattern that is of interest is at the data link level, and
268 the relationship between what you type and what the controllers trans‐
269 mit can be complicated.
270 This means that if you have a data-dependent problem you will probably
272 have to do a lot of testing to find it. If you are lucky, you may man‐
273 age to find a file that either can't be sent across your network or
274 that takes much longer to transfer than other similar length files.
275 You can then examine this file for repeated patterns that you can test
276 using the -p option of ping.
277
279 The TTL value of an IP packet represents the maximum number of IP
280 routers that the packet can go through before being thrown away. In
281 current practice you can expect each router in the Internet to decre‐
282 ment the TTL field by exactly one.
283 The TCP/IP specification states that the TTL field for TCP packets
285 should be set to 60, but many systems use smaller values (4.3 BSD uses
286 30, 4.2 used 15).
287 The maximum possible value of this field is 255, and most Unix systems
289 set the TTL field of ICMP ECHO_REQUEST packets to 255. This is why you
290 will find you can ``ping'' some hosts, but not reach them with tel‐
291 net(1) or ftp(1).
292 In normal operation ping prints the ttl value from the packet it
294 receives. When a remote system receives a ping packet, it can do one
295 of three things with the TTL field in its response:
296 · Not change it; this is what Berkeley Unix systems did before the
298 4.3BSD Tahoe release. In this case the TTL value in the received
299 packet will be 255 minus the number of routers in the round-trip
300 path.
301 · Set it to 255; this is what current Berkeley Unix systems do. In
303 this case the TTL value in the received packet will be 255 minus the
304 number of routers in the path from the remote system to the pinging
305 host.
306 · Set it to some other value. Some machines use the same value for ICMP
308 packets that they use for TCP packets, for example either 30 or 60.
309 Others may use completely wild values.
310
312 · Many Hosts and Gateways ignore the RECORD_ROUTE option.
313 · The maximum IP header length is too small for options like
315 RECORD_ROUTE to be completely useful. There's not much that that can
316 be done about this, however.
317 · Flood pinging is not recommended in general, and flood pinging the
319 broadcast address should only be done under very controlled condi‐
320 tions.
321
323 netstat(1), ifconfig(8).
324
326 The ping command appeared in 4.3BSD.
327 The version described here is its descendant specific to Linux.
329
331 ping requires CAP_NET_RAWIO capability to be executed. It may be used
332 as set-uid root.
333
335 ping is part of iputils package and the latest versions are available
336 in source form at http://www.skbuff.net/iputils/iputils-cur‐
337 rent.tar.bz2.
338iputils-100418 13 July 2010 PING(8)